Path planning method and device and autonomous vehicle

Abstract

The present disclosure provides a path planning method, device and automatic driving vehicle, and relates to the technical field of automatic driving and intelligent transportation. The path planning method comprises: determining a target road region available for a vehicle to perform a target driving behavior based on initial state information of the vehicle and obstacle perception information; obtaining a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior; generating an initial path for implementing the target driving behavior based on the initial state information and the reference path; and optimizing the initial path based on a preset optimization target under the constraints of the target road region and the state information range to generate a target path for implementing the target driving behavior.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular to the technical field of automatic driving and intelligent transportation, and more particularly to a path planning method and device, an electronic device, a computer readable storage medium, a computer program product and an automatic driving vehicle. BACKGROUND

[0002] Automatic driving technology involves multiple aspects such as environment perception, behavior decision, path planning and motion control. By relying on the collaborative work of sensors, vision computing systems and positioning systems, a vehicle with automatic driving function can automatically run without the operation of a driver or with only a small amount of operation of the driver.

[0003] Path planning in automatic driving refers to determining the best driving path of a vehicle on a road through an algorithm to achieve safe and efficient driving of the vehicle. The planned path is usually represented as a series of trajectory points. Therefore, in some cases, path planning can also be referred to as trajectory planning.

[0004] The methods described in this section can not necessarily be the methods previously conceived or used. Unless otherwise indicated, nothing in this section should be assumed to be prior art merely because of its inclusion in this section. Similarly, issues mentioned in this section should not be assumed to have been admitted to be prior art in any jurisdiction unless otherwise indicated. SUMMARY

[0005] The present disclosure provides a path planning method and device, an electronic device, a computer readable storage medium, a computer program product and an automatic driving vehicle.

[0006] According to an aspect of the present disclosure, a path planning method is provided, comprising: determining a target road region available for a vehicle to perform a target driving behavior based on initial state information of the vehicle and obstacle perception information; obtaining a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior; generating an initial path for implementing the target driving behavior based on the initial state information and the reference path; and under the constraints of the target road region and the state information range, optimizing the initial path based on a preset optimization target to generate a target path for implementing the target driving behavior.

[0007] According to an aspect of the present disclosure, a path planning apparatus is provided, comprising: a first determining module configured to determine, based on initial state information of a vehicle and obstacle perception information, a target road region available for the vehicle to perform a target driving behavior; a first obtaining module configured to obtain a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior; a first generating module configured to generate, based on the initial state information and the reference path, an initial path for implementing the target driving behavior; and a second generating module configured to optimize the initial path based on a preset optimization target under constraints of the target road region and the state information range, to generate a target path for implementing the target driving behavior.

[0008] According to an aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above method.

[0009] According to an aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, the computer instructions being used to cause a computer to perform the above method.

[0010] According to an aspect of the present disclosure, a computer program product is provided, comprising a computer program executable by a processor to implement the above method.

[0011] According to an aspect of the present disclosure, an autonomous vehicle is provided, comprising the above electronic device.

[0012] According to one or more embodiments of the present disclosure, the rationality and efficiency of path planning for complex driving behaviors can be improved.

[0013] It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become apparent through the following description. BRIEF DESCRIPTION OF DRAWINGS

[0014] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments and together with the description serve to explain exemplary implementations of the application. The illustrated embodiments are exemplary only and not limiting of the scope of the claims. In all the drawings, like reference numerals refer to like parts throughout the several views.

[0015] Figure 1A schematic diagram showing an exemplary system in which various methods described herein can be implemented according to embodiments of the present disclosure is shown;

[0016] Figure 2 A flowchart showing a path planning method according to embodiments of the present disclosure is shown;

[0017] Figure 3 A schematic diagram showing adjustment of road boundaries based on obstacle perception information according to embodiments of the present disclosure is shown;

[0018] Figure 4 A schematic diagram showing adjustment of road boundaries for a U-turn scenario according to embodiments of the present disclosure is shown;

[0019] Figure 5 A schematic diagram showing constraints of a target road region according to embodiments of the present disclosure is shown;

[0020] Figure 6 A schematic diagram showing an index back-off problem arising without an incremental point finding scheme according to embodiments of the present disclosure is shown;

[0021] Figure 7 A schematic diagram showing a path planning process according to embodiments of the present disclosure is shown;

[0022] Figure 8 A structural block diagram of a path planning apparatus according to embodiments of the present disclosure is shown; and

[0023] Figure 9 A structural block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0024] Exemplary embodiments of the present disclosure are described herein with reference to the accompanying drawings, which are incorporated in this specification, wherein various details of embodiments of the present disclosure are set forth in order to provide a thorough understanding of the present disclosure. It will be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described herein, embody the principles of the present disclosure and are included within its scope. As well, in order to clearly illustrate the present disclosure and to facilitate understanding of various aspects thereof, well-known functions or constructions are not described in detail.

[0025] In the present disclosure, the terms "first", "second", etc. are used to describe various elements only to distinguish one element from another, and are not intended to imply a relative position relationship, a temporal sequence relationship, or a relative importance between the elements, unless otherwise specified. In some examples, the first element and the second element can refer to the same instance of the element, and in some cases, based on the context of the description, they can also refer to different instances.

[0026] The terminology used in the description of the various described examples in the present disclosure is for the purpose of describing particular examples only and is not intended to be limiting. Unless specifically defined otherwise, an element that is a single can also be multiple and vice versa. Furthermore, the term "and / or" as used herein encompasses any and all possible combinations of the listed items. "Multiple" means two or more.

[0027] In the technical solutions of the present disclosure, the acquisition, storage and application of user personal information comply with relevant laws and regulations and do not violate public order and good customs.

[0028] There are some relatively complex driving behaviors in the driving process of an autonomous vehicle, such as turning around at an intersection, left turning, right turning, and reversing into a garage, and the like. The rationality and efficiency of path planning for complex driving behaviors are important prerequisites for ensuring the safe and efficient driving of an autonomous vehicle.

[0029] In the related art, a preset path can be used to implement complex driving behaviors. For example, a turning-around path, a left-turning path, a right-turning path, and the like at each intersection can be marked in advance in a map (which can be a high-definition map). An autonomous vehicle drives according to these pre-marked paths, thereby implementing turning-around, left-turning, right-turning, and the like. However, the pre-marked paths can not be consistent with the current actual application scenario of the autonomous vehicle, and the rationality and feasibility are poor. For example, due to a large number of pedestrians and vehicles at the intersection and complex traffic conditions, the pre-marked path can be blocked by obstacles at the intersection, causing the autonomous vehicle to be unable to successfully implement the specified driving behavior according to the path, and a remote driver needs to take over, which reduces the safety and driving efficiency of the autonomous vehicle.

[0030] In some other related art, instead of using a preset path, a real-time path planning can be performed based on the current application scenario of the autonomous vehicle to solve an optimal driving path. However, this way has a large amount of calculation and a long solving time, and even can not obtain a feasible solution, causing the autonomous vehicle to be unable to safely and efficiently complete the specified driving behavior.

[0031] To solve the above problems, the disclosure provides a path planning method. The method first combines the initial state information of the vehicle and the reference path to quickly generate an initial path that conforms to the reference path and is consistent with the actual state of the vehicle, and has strong feasibility. Further, the strong feasibility of the initial path is optimized in combination with the current road constraint and behavior state constraint, which can ensure the solution speed and quality of path optimization, so as to quickly obtain a reasonable and feasible target path, improve the rationality and efficiency of path planning for complex driving behaviors, and enable the vehicle to safely and efficiently automatically travel.

[0032] Embodiments of the disclosure will be described in detail below with reference to the accompanying drawings.

[0033] Figure 1 A schematic diagram of an example system 100 in which various methods and apparatus described herein can be implemented according to embodiments of the disclosure is shown. Referring to Figure 1 The system 100 includes a motor vehicle 110, a server 120, and one or more communication networks 130 coupling the motor vehicle 110 to the server 120.

[0034] In embodiments of the disclosure, the motor vehicle 110 can include an electronic device according to embodiments of the disclosure and / or be configured to perform methods according to embodiments of the disclosure.

[0035] The server 120 can run one or more services or software applications that enable the performance of methods of embodiments of the disclosure. In certain embodiments, the server 120 can also provide other services or software applications, which can include non-virtual and virtual environments. In Figure 1 In the configuration shown, the server 120 can include one or more components that implement the functionality performed by the server 120. These components can include software components, hardware components, or a combination thereof, executable by one or more processors. Users of the motor vehicle 110 can in turn utilize one or more client applications to interact with the server 120 to utilize the services provided by these components. It should be understood that various different system configurations are possible, which can differ from the system 100. Therefore, Figure 1 The system 100 is one example of a system for implementing the various methods described herein and is not intended to be limiting.

[0036] The server 120 can include one or more general purpose computers, special purpose server computers (e.g., PC (personal computer) servers, UNIX servers, mid-range servers), blade servers, mainframe computers, server clusters, or any other appropriate arrangement and / or combination. The server 120 can include one or more virtual machines running virtual operating systems, or other computing architectures involving virtualization (e.g., one or more flexible pools of logical storage devices that can be virtualized to maintain virtual storage devices for the server). In various embodiments, the server 120 can run one or more services or software applications that provide the functionality described below.

[0037] The computing units in the server 120 can run one or more operating systems including any of the operating systems described above, as well as any commercially available server operating systems. The server 120 can also run any of a variety of additional server applications and / or mid-tier applications, including HTTP servers, FTP servers, CGI servers, JAVA servers, database servers, etc.

[0038] In some embodiments, the server 120 can include one or more applications to analyze and consolidate data feeds and / or event updates received from the motor vehicles 110. The server 120 can also include one or more applications to display the data feed and / or real-time events via one or more display devices of the motor vehicles 110.

[0039] The network 130 can be any type of network familiar to those skilled in the art that can support data communications using any of a variety of available protocols, including without limitation TCP / IP, SNA, IPX, etc. As an example, one or more of the networks 130 can be a satellite communications network, a local area network (LAN), an Ethernet-based network, a token ring, a wide area network (WAN), the Internet, a virtual network, a virtual private network (VPN), an intranet, an extranet, a blockchain network, a public switched telephone network (PSTN), an infrared network, a wireless network (including, for example, a Bluetooth, Wi-Fi, and / or ZigBee network), and / or any combination of these and other networks.

[0040] The system 100 can also include one or more databases 150. In certain embodiments, these databases can be used to store data and other information. For example, one or more of the databases 150 can be used to store information such as audio files and video files. The databases 150 can reside in various locations. For example, a data store used by the server 120 can be local to the server 120, or can be remote from the server 120 and can communicate with the server 120 via a network- or dedicated-based connection. The databases 150 can be of different types. In certain embodiments, the databases 150 can be relational databases. One or more of these databases can store, update, and retrieve data to and from the databases in response to commands.

[0041] In certain embodiments, one or more of the databases 150 can also be used by the applications to store application data. The databases used by the applications can be different types of databases, such as key-value stores, object stores, or regular stores supported by a file system.

[0042] The motor vehicle 110 can include sensors 111 for perceiving the surrounding environment. The sensors 111 can include one or more of the following sensors: visual camera, infrared camera, ultrasonic sensor, millimeter wave radar, and laser radar. Different sensors can provide different detection accuracy and range. The camera can be installed at the front, rear, or other positions of the vehicle. The visual camera can capture the situation inside and outside the vehicle in real time and present it to the driver and / or passenger. In addition, through analysis of the pictures captured by the visual camera, information such as traffic signal indication, intersection situation, running state of other vehicles, etc. can be obtained. The infrared camera can capture objects in night vision conditions. The ultrasonic sensor can be installed around the vehicle to measure the distance of objects outside the vehicle from the vehicle by taking advantage of the strong directivity of ultrasonic waves. The millimeter wave radar can be installed at the front, rear, or other positions of the vehicle to measure the distance of objects outside the vehicle from the vehicle by taking advantage of the characteristics of electromagnetic waves. The laser radar can be installed at the front, rear, or other positions of the vehicle to detect object edges, shape information, and thus perform object recognition and tracking. Due to the Doppler effect, the radar device can also measure the speed change of the vehicle and moving objects.

[0043] The motor vehicle 110 can further comprise a communication device 112. The communication device 112 can comprise a satellite positioning module capable of receiving satellite positioning signals (e.g. Beidou, GPS, GLONASS and GALILEO) from satellites 141 and generating coordinates based on these signals. The communication device 112 can further comprise a module for communicating with mobile communication base stations 142, which can implement any suitable communication technology, such as GSM / GPRS, CDMA, LTE, etc. current or evolving wireless communication technologies (e.g. 5G technology). The communication device 112 can further have a vehicle-to-everything (V2X) module configured for enabling vehicle-to-vehicle (V2V) communication with other vehicles 143 and vehicle-to-infrastructure (V2I) communication with infrastructure 144, for example. In addition, the communication device 112 can further have a module configured to communicate with user terminals 145 (including but not limited to smartphones, tablets or wearable devices such as watches) for example through a wireless local area network using IEEE 802.11 standards or Bluetooth. With the communication device 112, the motor vehicle 110 can further access the server 120 via the network 130.

[0044] The motor vehicle 110 can further comprise an inertial navigation module. The inertial navigation module can be combined with the satellite positioning module into a combined positioning system for enabling initial positioning of the motor vehicle 110.

[0045] The motor vehicle 110 can further comprise a control device 113. The control device 113 can comprise a processor, such as a central processing unit (CPU) or a graphics processing unit (GPU), or other special purpose processor, in communication with various types of computer readable storage devices or media. The control device 113 can comprise an autonomous driving system for automatically controlling various actuators in the vehicle, accordingly the motor vehicle 110 is an autonomous vehicle. The autonomous driving system is configured to control the powertrain, steering system and braking system of the motor vehicle 110 (not shown) via a plurality of actuators to control acceleration, steering and braking, respectively, in response to inputs from a plurality of sensors 111 or other input devices, without or with limited human intervention. Part of the processing functions of the control device 113 can be implemented through cloud computing. For example, certain processing can be performed using an on-board processor, while other processing can be performed using computing resources of the cloud. The control device 113 can be configured to perform the methods according to the present disclosure. In addition, the control device 113 can be implemented as one example of an electronic device on the motor vehicle side (client) according to the present disclosure.

[0046] Figure 1 The system 100 can be configured and operated in various ways to enable the application of various methods and devices described according to the present disclosure. For example, the motor vehicle 110 or the server 120 can be configured to perform the path planning method 200 of embodiments of the present disclosure to generate a target path for realizing a target driving behavior. In turn, the motor vehicle 110 is controlled to travel according to the target path, thereby realizing the target driving behavior.

[0047] Figure 2 A flowchart of the path planning method 200 according to embodiments of the present disclosure is shown. As described above, the execution subject of the method 200 can be a motor vehicle (for example, the motor vehicle 110 shown in Figure 1 ) or a server (for example, the server 120 shown in Figure 1 ).

[0048] As shown in Figure 2 , the method 200 includes steps S210-S240.

[0049] In step S210, a target road region available for the vehicle to perform a target driving behavior is determined based on initial state information of the vehicle and obstacle perception information.

[0050] In step S220, a reference path for realizing the target driving behavior and a state information range corresponding to the target driving behavior are obtained. The state information range indicates a value range of the state information of the vehicle when performing the target driving behavior.

[0051] In step S230, an initial path for realizing the target driving behavior is generated based on the initial state information and the reference path.

[0052] In step S240, the initial path is optimized based on a preset optimization target under the constraints of the target road region and the state information range to generate a target path for realizing the target driving behavior.

[0053] According to embodiments of the present disclosure, first, the initial state information of the vehicle and the reference path are combined to quickly generate an initial path that fits the reference path and is consistent with the actual state of the vehicle and has strong feasibility. In turn, the strong feasible initial path is optimized in combination with the current road constraints and behavior state constraints, which can ensure the solving speed and quality of the path optimization, thereby quickly obtaining a reasonable and feasible target path, improving the rationality and efficiency of the path planning for complex driving behaviors, and enabling the vehicle to safely and efficiently automatically travel.

[0054] The method 200 of the embodiments of the present disclosure can be performed in a Cartesian coordinate system. Compared with the scheme of performing path planning in a Frenet coordinate system in the related art, the path planning in the Cartesian coordinate system does not have projection loss and is more versatile, and is more suitable for planning of a path with large curvature.

[0055] In step S210, a target road region available for the vehicle to perform a target driving behavior is determined based on initial state information of the vehicle and obstacle perception information.

[0056] In the embodiments of the present disclosure, the target driving behavior is a relatively complex driving behavior for the vehicle. The complexity can be embodied in multiple aspects such as environmental interaction and vehicle control. For example, the driving behavior at a road junction can be taken as the target driving behavior, because the vehicle needs to make a decision in combination with the surrounding environment (such as traffic lights, pedestrians, other vehicles, etc.) when driving at the road junction. For example, the driving behavior at the road junction, such as U-turn, left turn, right turn, etc. can be taken as the target driving behavior. For another example, the driving behavior with large change in direction usually requires complex and fine control of the vehicle, and therefore the driving behavior with large change in direction, such as U-turn, reversing into a garage, etc. can be taken as the target driving behavior.

[0057] According to some embodiments, a target driving behavior set can be pre-set. The target driving behavior set includes multiple target driving behaviors. Accordingly, whether the current driving behavior of the vehicle is the target driving behavior can be determined by judging whether the current driving behavior of the vehicle is a driving behavior in the set.

[0058] The method 200 of the embodiments of the present disclosure can be a local path planning method. For example, a navigation path from a starting point to a destination can be planned by using a navigation planning algorithm. The navigation path can guide the vehicle to generate a series of driving behaviors, such as starting from point A, driving straight along lane B, making a U-turn at road junction C to enter lane D, driving straight along lane D, making a right turn at road junction E to enter lane F, and driving straight along lane F to reach destination G. It can be understood that different segments in the navigation path correspond to different driving behaviors. According to some embodiments, the navigation path can be segmented to obtain multiple path segments, and the driving behavior corresponding to each path segment can be determined. Accordingly, when the vehicle starts to drive along a certain path segment, it can be considered that the vehicle starts to perform the driving behavior corresponding to the path segment. If the driving behavior is a relatively complex target driving behavior, the method 200 of the embodiments of the present disclosure is used to re-plan the path segment to obtain an optimized path segment. In this way, fine planning of a local path in a global navigation path is realized, and it is ensured that the vehicle can safely and smoothly perform the target driving behavior, thereby ensuring that the vehicle can safely and efficiently drive.

[0059] The method 200 of the embodiments of the present disclosure can perform path planning based on the current state information (i.e., initial state information) of the vehicle and the obstacle perception information when the vehicle just starts to perform the target driving behavior, to generate an optimized target path for implementing the target driving behavior.

[0060] The state information of the vehicle may, for example, include the position (represented by two-dimensional coordinates (x, y) in a Cartesian coordinate system), speed, acceleration, direction angle (heading angle), curvature of the lane where the vehicle is located, and the like of the vehicle.

[0061] The obstacle perception information may, for example, include the position, size, speed, orientation, and the like of the obstacles in the surrounding environment. According to some embodiments, the obstacle perception information may, for example, be obtained by inputting the environmental image or point cloud data collected by the vehicle into a trained object recognition model.

[0062] Based on the initial state information of the vehicle and the obstacle perception information, a target road region available for performing the target driving behavior can be determined.

[0063] According to some embodiments, as described above, the initial state information of the vehicle may, for example, include position information. Accordingly, the target road region can be determined through steps S211 and S212.

[0064] In step S211, based on the position information of the vehicle, an initial road region corresponding to the target driving behavior is determined.

[0065] In step S212, based on the obstacle perception information, the boundary of the initial road region is adjusted to obtain the target road region.

[0066] According to the above embodiments, the target road region available for the vehicle to perform the target driving behavior is determined based on the current obstacle perception information of the vehicle, which can improve the rationality and feasibility of subsequent path planning.

[0067] According to some embodiments, the position information of the vehicle may, for example, be two-dimensional coordinates or three-dimensional coordinates in a world coordinate system. To simplify the calculation, two-dimensional coordinates, such as longitude and latitude coordinates, may, for example, be used.

[0068] According to some embodiments, in step S211, based on the position information of the vehicle, the lane where the vehicle is currently located can be obtained from the high-definition map, and the road region used by the vehicle when performing the target driving behavior is obtained according to the lane as the initial road region. The initial road region may, for example, include the lane region where the vehicle is currently located, the intersection region connected to the lane, and the like.

[0069] In step S212, by adjusting the boundary of the initial road region based on the obstacle perception information, the target road region that is more consistent with the current driving scenario of the vehicle can be obtained.

[0070] According to some embodiments, the obstacle perception information can comprise a plurality of vertex coordinates of the obstacle. The plurality of vertex coordinates of the obstacle can be obtained by processing an environment image or point cloud data collected by the vehicle using a trained object recognition model, for example. By connecting the plurality of vertex coordinates of the obstacle, an envelope line of the obstacle can be obtained. The envelope line can represent the position and size of the obstacle.

[0071] In the case where the obstacle perception information comprises the plurality of vertex coordinates of the obstacle, step S212 can comprise steps S2121-S2123.

[0072] In step S2121, in response to at least one vertex coordinate in the plurality of vertex coordinates being located within the initial road region, a contour line of the obstacle within the initial road region is determined based on the at least one vertex coordinate.

[0073] In step S2122, a first boundary segment intersecting the contour line and a second boundary segment of a preset length adjacent to the first boundary segment are determined as a target boundary segment.

[0074] In step S2123, the target boundary segment is adjusted towards the inside of the initial road region, so that the adjusted target boundary segment bypasses the contour line.

[0075] According to the above embodiments, by adjusting the boundary of the initial road region based on the position information (vertex coordinates) of the obstacle, a target road region is obtained. The target road region is the actual road region available for the vehicle to travel. By performing path planning in the target road region, the rationality and feasibility of the planned path can be improved, thereby ensuring the safe and efficient travel of the vehicle.

[0076] According to some embodiments, in step S2121, it can be determined whether each vertex coordinate of the obstacle is located within the initial road region respectively. If at least one vertex coordinate is located within the initial road region, the obstacle at least partially intrudes into the road boundary, and thus affects the behavior of the vehicle. By connecting the at least one vertex coordinate located within the initial road region, a contour line of the obstacle within the initial road region can be obtained.

[0077] The first boundary segment is a boundary segment located between the two intersection points of the contour line and the road boundary. The second boundary segment is a boundary segment of a preset length located on both sides of the first segment. The target boundary segment is a boundary segment composed of the first boundary segment and two second boundary segments.

[0078] According to some embodiments, in step S2123, each vertex coordinate located in the initial road region can be translated by a preset distance (e.g., 0.2 meters, 0.5 meters, etc.) to the inside of the initial road region respectively, to obtain a translated vertex coordinate. By connecting the two end points of the target boundary segment and the translated vertex coordinates in turn, an adjusted target boundary segment is obtained.

[0079] Figure 3 A schematic diagram of adjusting the boundary 320 of the initial road region 300 based on obstacle perception information according to an embodiment of the present disclosure is shown. The initial road region 300 can be a lane region where the vehicle is currently located, and the center line 310 of the lane is a dashed line.

[0080] The vehicle perceives the surrounding environment to obtain 4 vertexes A1, A2, A3, A4 of the obstacle A and 5 vertexes B1, B2, B3, B4, B5 of the obstacle B.

[0081] For the obstacle A, the distance of each vertex A1-A4 to the lane center line 310 can be calculated respectively. If the distance of a vertex to the lane center line 310 is less than half of the lane width (which can be obtained from the map), it is determined that the vertex is located in the initial road region 300. Thus, the vertexes A3 and A4 located in the inside of the initial road region 300 are obtained, and further the contour line C1A4A3C2 of the obstacle A inside the initial road region 320 is obtained. The intersection points of the contour line and the road boundary 320 are C1 and C2. The upper intersection point C1 is translated by a distance (e.g., 0.5 meters) upwards along the road boundary 320 to obtain point C1’, and the lower intersection point C2 is translated by a distance (e.g., 0.5 meters) downwards along the road boundary 320 to obtain point C2’, and further the target boundary segment C1’C2’ is obtained. The target boundary segment C1’C2’ is adjusted to the inside of the road to bypass the contour line C1A4A3C2 of the obstacle A inside the road. Specifically, the vertexes A3 and A4 can be translated by a distance (e.g., 0.5 meters) to the inside of the road respectively to obtain vertexes A3’ and A4’. The points C1’, A4’, A3’, C2’ are connected in turn to obtain the adjusted target boundary segment C1’A4’A3’C2’.

[0082] Similarly, for the obstacle B, the contour line C3B1B2B3C4 of the obstacle B located in the initial road region 320 can be obtained. The points C3 and C4 are translated by a distance (e.g., 0.5 meters) upwards and downwards along the road boundary respectively to obtain the target boundary segment C3’C4’. The vertexes B1-B3 are translated by a distance (e.g., 0.5 meters) to the inside of the road respectively to obtain vertexes B1’-B3’. The points C3’, B1’, B2’, B3’, C4’ are connected in turn to obtain the adjusted target boundary segment C3’B1’B2’B3’C4’.

[0083] According to some embodiments, in the scenario of vehicle U-turn, in addition to the initial road region can be adjusted based on the obstacle perception information, the initial road region can also be adjusted based on the characteristics of the U-turn behavior, so that the planned path can be feasible. For example, in the U-turn behavior, the vehicle can exceed the range of the lane in the entry curve part (i.e., the connection between the lane before the U-turn and the intersection) or the exit curve part (i.e., the connection between the intersection and the lane after the U-turn), so that the driving path of the vehicle is expanded outward compared to the lane center line. In the case of a relatively narrow U-turn lane (i.e., the width of the lane is less than a threshold), if the outward expansion occurs in the entry curve part of the U-turn path, the state constraint and the road boundary constraint can conflict when solving the optimization problem, resulting in no feasible solution. Therefore, in this case, the road right boundary of the exit curve part can be expanded outward, so that it is more likely to plan a U-turn path that expands outward in the exit curve part, improving the solving effect of the optimization problem of narrow roads.

[0084] Figure 4 A schematic diagram of adjusting the road boundary for the U-turn scenario according to an embodiment of the present disclosure is shown. As shown, the left and right boundaries of the initial road region are 410 and 420, respectively. In the case of a relatively narrow lane, the road right boundary 422 of the exit curve part is preferably expanded to obtain the boundary 424, so as to prevent the vehicle from expanding outward in the entry curve part. Figure 4

[0085] In step S220, a reference path for implementing the target driving behavior is obtained.

[0086] In an embodiment of the present disclosure, the reference path can be understood as a reference line for path planning of the target driving behavior. The reference path is composed of a series of trajectory points, each trajectory point having position, speed, direction, etc. In an embodiment of the present disclosure, the trajectory points in the reference path are denoted as first trajectory points.

[0087] According to some embodiments, the reference path can be obtained by any of the following operations: determining a feasible path obtained by searching in the target road region using a search algorithm as the reference path; inputting at least the information of the target road region into a trained machine learning model to obtain the reference path output by the machine learning model; or obtaining a reference path marked in a map.

[0088] According to some embodiments, a feasible path that can avoid obstacles can be searched out using the A* search algorithm using the map and the obstacle perception information, as the reference path.

[0089] ​According to some embodiments, information of a target road region, such as width, curvature, position coordinates of one or more boundary points, etc., can be input into a trained machine learning model to obtain a reference path output by the machine learning model. The machine learning model may, for example, be a neural network model. The machine learning model may, for example, be trained with information of sample road regions and their corresponding reference path labels as samples.

[0090] According to some embodiments, a center line of each lane is marked in the map. Accordingly, the lane center line of the current lane can be taken as the reference path.

[0091] In step S220, a state information range corresponding to the target driving behavior is also obtained. The state information range indicates a range of values of state information of the vehicle when performing the target driving behavior.

[0092] According to some embodiments, the state information range can be preset. For example, for the behavior of turning around, the speed of the vehicle is usually required to be no more than 15 km / h; for the behaviors of left turning and right turning, the speed of the vehicle is usually required to be no more than 30 km / h.

[0093] In step S230, an initial path for implementing the target driving behavior is generated based on the initial state information and the reference path.

[0094] The reference path can be understood as a planned path of the vehicle in an ideal state. The initial state information describes the actual state of the vehicle. The initial path is generated in combination with the initial state information and the reference path, so that the generated initial path can not only fit the reference path, but also be consistent with the actual state of the vehicle, and has strong feasibility.

[0095] According to some embodiments, in step S230, the initial path can be generated by using a preset target PID controller based on the initial state information of the vehicle and the reference path.

[0096] The PID controller is a Proportion Integration Differentiation controller. According to the above embodiments, by using the PID controller, an initial path with strong feasibility, which can fit the reference path and be consistent with the actual state of the vehicle, can be quickly generated.

[0097] According to some embodiments, the reference path is represented as a first sequence of trajectory point information, and the first sequence of trajectory point information includes first state information of each of a plurality of first trajectory points. The initial path is represented as a second sequence of trajectory point information, and the second sequence of trajectory point information includes second state information of each of a plurality of second trajectory points. The first state information and the second state information may, for example, include two-dimensional position coordinates in a Cartesian coordinate system, speed, acceleration, direction angle, road curvature, etc.

[0098] Accordingly, the step S230 can include: sequentially performing the following operations S231-S234 for each first trajectory point in the plurality of first trajectory points, to determine the second state information of the second trajectory point corresponding to each first trajectory point.

[0099] In operation S231, the first state information of the first trajectory point is determined as the expected state information.

[0100] In operation S232, the current state information is obtained, wherein, in response to the first trajectory point being the first first trajectory point in the plurality of first trajectory points, the initial state information is determined as the current state information, and in response to the first trajectory point being any first trajectory point other than the first first trajectory point, the second state information of the second trajectory point corresponding to the previous first trajectory point is determined as the current state information.

[0101] In operation S233, the control information is generated by the target PID controller based on the deviation between the current state information and the expected state information.

[0102] In operation S234, the second state information of the second trajectory point corresponding to the first trajectory point is determined based on the control information.

[0103] According to the above embodiments, the state information (i.e., the second state information) of each trajectory point (i.e., the second trajectory point) in the initial path can be generated one by one by using the target PID controller.

[0104] According to some embodiments, the operation S234 can include operations S2341 and S2342.

[0105] In operation S2341, the state change information is determined based on the control information.

[0106] In operation S2342, the current state information is superimposed with the state change information to obtain the second state information of the second trajectory point corresponding to the first trajectory point.

[0107] According to some embodiments, the control information includes a plurality of control information items, and the state change information includes a plurality of change information items, each change information item in the plurality of change information items being obtained by integrating at least one control information item in the plurality of control information items.

[0108] According to the above embodiments, the state change information is obtained by integrating the control information, and the integration operation makes the state change information more continuous and smooth, thereby ensuring the continuity and smoothness of the initial trajectory and improving the stability of the vehicle during driving and the comfort of the passengers.

[0109] An example of generating an initial path using a target PID controller is given below.

[0110] When performing path planning, a six-degree-of-freedom model can be used to model the vehicle. The six-degree-of-freedom model uses a Cartesian coordinate system, and there is no loss of projection. Although it has more optimization variables and is more difficult to solve than the path planning problem in the Frenet coordinate system, the use of an iLQR (Iterative Linear Quadratic Regulator) solver can reduce the difficulty of solving and improve the speed of solving, meeting the real-time requirements of path planning.

[0111] The six degrees of freedom in the six-degree-of-freedom model of the vehicle are six state information of the vehicle, including two position coordinates (x, y) in the Cartesian coordinate system, velocity v, acceleration a, orientation angle θ, and road curvature κ. The six state information satisfies the following state transition equation:

[0112]

[0113] where j and τ are control information output by the target PID controller. j is the rate of change of acceleration a, and τ is the rate of change of curvature κ. Taking j and τ as control information makes the above six state information able to be obtained by integrating the control information. This can ensure the continuity and smoothness of the initial trajectory, and improve the stability of the vehicle during driving and the comfort of passengers.

[0114] According to the above embodiment, the state transition equation of the vehicle kinematic model can be fully utilized to accelerate the solution, strictly guarantee the kinematic constraints, and the solution efficiency is higher than that of the optimization problem solving method using the kinematic model as a constraint. At the same time, the nonlinear programming problem solver iLQR is directly used to solve the path planning problem, and there is no linearization error, so the solution accuracy can be guaranteed.

[0115] The target PID controller takes the velocity error e v , the orientation error e θ , and the position error e pos as inputs:

[0116] e v =v-v ref (7)

[0117] e θ =θ-θ ref (8)

[0118]

[0119] In the above formula, V ref , θ ref , (xref , y ref ) are the velocity, the direction angle and the position coordinates of the reference trajectory point (i.e., the first trajectory point) in the reference path, respectively, v, θ, (x, y) are the velocity, the direction angle and the position coordinates of the second trajectory point in the initial trajectory corresponding to the reference trajectory point, respectively. d represents the offset direction of the second trajectory point relative to the first trajectory point, d = -1 if the second trajectory point is located at the left side of the first trajectory point, and d = 1 if the second trajectory point is located at the right side of the first trajectory point.

[0120] The target PID controller performs the following calculation based on its own PID parameters:

[0121]

[0122]

[0123] wherein, are the PID parameters of the target PID controller (are preset constants).

[0124] Based on the three intermediate information u v , u θ , u pos obtained by formula (10)-(12), two control information are output:

[0125] u[0] = u v = j (13)

[0126] u[1] = u θ + u pos = τ (14)

[0127] Based on the above state transition equation, the state change information of the six degrees of freedom can be calculated by integrating the control information u[0], u[1]. The current state information is superimposed with the state change information to obtain the updated state information.

[0128] According to some embodiments, the method 200 further comprises steps S251-S253. Steps S251-S253 are used to select a candidate PID controller with the best control effect from a plurality of candidate PID controllers with different PID parameters as the target PID controller.

[0129] In step S251, a plurality of candidate PID controllers are obtained.

[0130] In step S252, for any candidate PID controller in the plurality of candidate PID controllers: based on the test initial state information and the test reference path, a test initial path is generated using the candidate PID controller; and the difference between the test initial path and the test reference path is calculated.

[0131] In step S253, the candidate PID controller with the minimum difference is determined as the target PID controller.

[0132] According to the above embodiments, the optimal candidate PID controller is selected from a plurality of candidate PID controllers with different parameters, thereby ensuring the planning effect of the initial path.

[0133] According to some embodiments, in step S251, the PID parameters of the plurality of candidate PID controllers are different. The PID parameters of each candidate PID controller can be set in various ways, such as random sampling within a preset sampling range, uniform sampling at a fixed interval within a preset sampling range, or generated by a preset generation function.

[0134] According to some embodiments, in step S252, for each candidate PID controller, different initial state information can be used to test the initial path generation effect of the candidate PID controller. The initial state information used to test the candidate PID controller is denoted as test initial state information. The test initial state information includes test position (x0, y0), test speed v0, test acceleration a0, test direction angle θ0, and test curvature κ0. Considering different driving scenarios, the test initial state information can be set with corresponding constraints. For example, in the turning scenario, the driving speed of the vehicle is usually small, the vehicle usually completes a 180° turn, and the curvature of the entry / exit lane is usually small, so the sampling ranges of v0, θ0, and κ0 can be set as follows, where the unit of v0 is km / h:

[0135] v0∈[5, 15] (15)

[0136] θ0∈[-π / 2, π / 2] (16)

[0137] κ0∈[-0.2, 0.2] (17)

[0138] For the remaining three test initial state information (x0, y0) and a0, since the vehicle can theoretically be at any position, and there is no special requirement for the value of a0 for vehicle turning, the sampling ranges of x0, y0, and a0 are not limited.

[0139] In step S252, each candidate PID controller can be tested by using multiple sets of test initial state information. Specifically, for each set of test initial state information, a corresponding test initial path can be generated by using the candidate PID controller based on the set of test initial state information and the preset test reference path. A difference between each test initial path and the test reference path can be calculated. The difference can be, for example, a mean square error, i.e., a sum of squares of distances between position coordinates of each trajectory point in the test initial path and position coordinates of a corresponding trajectory point in the test reference path. Further, an average of the differences corresponding to the multiple sets of test initial state information, i.e., an average difference, can be calculated.

[0140] In step S253, the candidate PID controller with the smallest average difference can be determined as the target PID controller.

[0141] According to some embodiments, if, in a test initial path generated by a certain candidate PID controller, an absolute value of a direction angle of a certain trajectory point is greater than π / 2, i.e., |θ|>π / 2, the candidate PID controller can be considered unstable and excluded from the candidate range of the target PID controller. That is, the candidate PID controller is not used as the target PID controller.

[0142] After the initial path is generated in step S230, in step S240, the initial path is optimized based on a preset optimization objective under the constraints of the target road region and the state information range to generate a target path for implementing the target driving behavior.

[0143] According to some embodiments, step S240 can include steps S241 and S242.

[0144] In step S241, a loss value of the initial path is calculated based on a preset loss function.

[0145] In step S242, the initial path is adjusted under the constraints of the target road region and the state information range to reduce the loss value, and a target path is obtained.

[0146] According to the above embodiments, the initial path with relatively strong feasibility is optimized in combination with the current loss value, road constraints, and behavior state constraints, which can ensure the solving speed and quality of path optimization, and thus a reasonable and feasible target path can be quickly obtained.

[0147] It should be noted that in step S242, the initial path can be adjusted multiple times, i.e., the initial path is optimized multiple times until a preset termination condition is met. The preset termination condition can be, for example, that the number of adjustments reaches a threshold value (e.g., 150 times), the loss value is less than a threshold value, the loss value converges, the gradient of the loss value disappears, etc.

[0148] It can be understood that in the case of multiple optimizations of the initial path, the starting point of the later optimization is the result of the previous optimization. For example, the first optimization takes the initial path p1 as the starting point to obtain the optimized path p2; the second optimization takes the result p2 of the first optimization as the starting point to obtain the optimized path p3; the third optimization takes the result p3 of the second optimization as the starting point to obtain the optimized path p4; and so on, until the preset termination condition is reached, to obtain the target path.

[0149] It should be noted that in the embodiments of the present disclosure, the optimization of the initial path refers to the optimization of the second state information of each second trajectory point in the initial path. By optimizing the second state information of each second trajectory point, the loss value of the optimized path is continuously reduced until the optimal target path is obtained.

[0150] As described above, the initial path includes a plurality of second trajectory points. Accordingly, in step S241, the loss value of the initial path can be the sum of the single-point loss values of each second trajectory point. In step S242, each second trajectory point needs to satisfy the constraints of the target road region and the state information range. In order to simplify the description, the loss function and the constraints are introduced from the perspective of a single second trajectory point in the following.

[0151] According to some embodiments, the loss function can include a first loss term for indicating the difference between the initial path and the reference path and a second loss term for indicating the smoothness of the initial path.

[0152] According to the above-mentioned embodiments, the path optimization is performed in combination with the first loss term and the second loss term, which can ensure the safety and smoothness of the generated target path.

[0153] According to some embodiments, the calculation formula of the first loss term cost1 is as follows:

[0154]

[0155] where (x ref , y ref ), θ ref are the position coordinates and the direction angle of the first trajectory point in the reference path, respectively, and (x, y), θ are the position coordinates and the direction angle of the second trajectory point, respectively. w p , w h are the weights of the position and the direction, respectively, which can be preset constants.

[0156] According to some embodiments, the calculation formula of the second loss term cost2 can be as follows:

[0157]

[0158] wherein τ, κ, j, a are the rate of change of curvature, curvature, rate of change of acceleration, acceleration of the second trajectory point respectively. The first two terms in equation (19) represent the smoothness in lateral direction, and the last two terms represent the smoothness in longitudinal direction.

[0159] The loss function can be, for example, a weighted sum of the first loss term cost1 and the second loss term cost2. In the case where the weights of both are 1, the expression of the loss function is as follows:

[0160] cost = cost1 + cost2 (20)

[0161] According to some embodiments, in addition to the first loss term and the second loss term described above, the loss function can further include a third loss term for indicating whether the initial path meets the expected speed.

[0162] According to some embodiments, the calculation formula of the third loss term cost3 can be as follows:

[0163]

[0164] wherein v is the speed of the second trajectory point, v ref is the expected speed, w v is the weight of the speed. w v may be a preset constant. The expected speed can be set according to the expected speed of the target driving behavior. Taking the U-turn behavior as an example, the expected speed can be set as 4 m / s.

[0165] The loss function can be, for example, a weighted sum of the first loss term cost1, the second loss term cost2 and the third loss term cost3 described above. In the case where the weights of the three are 1, the expression of the loss function is as follows:

[0166] cost = cost1 + cost2 + cost3 (22)

[0167] According to some embodiments, the constraints of the state information range are as follows:

[0168] κ < κ max (23)

[0169] κ min < κ (24)

[0170] v < v max (25)

[0171] v min < v (26)

[0172] a < a max (27)

[0173] amin

[0174] wherein κ max , κ min are the maximum and minimum values of the curvature, v max , v min are the maximum and minimum values of the speed, a max , a min are the maximum and minimum values of the acceleration. The curvature κ, the speed v and the acceleration a of each second trajectory point resulting from the planning need to satisfy the above constraints.

[0175] Further, constraints can also be imposed on the control information j and τ, i.e. imposing control constraints:

[0176] j < j max (29)

[0177] j min < j (30)

[0178] τ < τ max (31)

[0179] τ min < τ (32)

[0180] wherein j max , j min are the maximum and minimum values of the rate of change of the acceleration, τ max , τ min are the maximum and minimum values of the rate of change of the curvature. In the optimization process of step S240, for each second trajectory point resulting from the planning, the corresponding control information j and τ can be calculated by the above state transition equations (5) and (6). The j and τ corresponding to each second trajectory point need to satisfy the above constraints.

[0181] The constraint of the target road region is used to constrain the position coordinates (x, y) of each second trajectory point resulting from the planning. According to some embodiments, the target road region can be represented by a series of road boundary points. It can be understood that, for each second trajectory point resulting from the planning, only the road boundary points close to the second trajectory point have a constraint effect on the second trajectory point. Therefore, in order to simplify the calculation, the target boundary points having a constraint effect on the current second trajectory point can be determined from the plurality of road boundary points constituting the target road region, and then the second trajectory point is adjusted and optimized under the constraint of the target boundary points.

[0182] Figure 5 A schematic diagram of the constraint of the target road region according to an embodiment of the present disclosure is shown. Figure 5 ​The solid black lines on the left and right sides mark the boundaries of the target road area, which consists of a series of road boundary points C. Rectangle 510 shows the vehicle's current position, i.e., the position of the second trajectory point. The midpoint of the vehicle's front axle (x...) is used as the reference point. f y f Using the center point (x) as the boundary, generate an envelope circle 520 that surrounds the front edge of the vehicle. Using the midpoint of the rear axle (x) as the boundary, generate an envelope circle 520 that surrounds the front edge of the vehicle. r y r Using the vehicle's envelope circle 520 and 530 as the center, an envelope circle 530 is generated that can surround the rear edge of the vehicle. The radii of the envelope circles 520 and 530 can be set as needed, as long as all four vertices of the rectangle 510 are located inside the envelope circle. For each road boundary point C, the projection n of the distance d from the road boundary point C to the center of each envelope circle in the normal vector direction can be calculated. If n < R, then the road boundary point is close to the vehicle and has a constraint effect on the vehicle, and the road boundary point can be determined as the target boundary point. According to this embodiment, using the vehicle's envelope circles 520 and 530, target boundary points that have a constraint effect on the currently planned second trajectory points can be quickly selected.

[0183] According to some embodiments, rule constraints related to the target driving behavior can be further set. For example, for U-turn behavior, an "incremental point finding" constraint can be set, that is, when solving the optimization problem, it is necessary to ensure that the index of the first trajectory point corresponding to the next second trajectory point is greater than the index of the first trajectory point corresponding to the previous second trajectory point, so as to avoid planning an unreasonable and infeasible path due to index rollback. Under the constraint of incremental point finding, the first trajectory point closest to the second trajectory point is taken as the first trajectory point that the second trajectory point needs to refer to when planning. The position, direction angle and velocity of the first trajectory point are (x) in the above formulas (18) and (21). ref y ref ), θ ref V ref .

[0184] Figure 6 A schematic diagram illustrates the index rollback problem that occurs when the incremental point-finding scheme of this disclosure is not employed. For example... Figure 6 As shown, the black line 610 represents the road boundary of the target road area, the green line 620 is the reference path, and the red line 630 is the path generated through the optimization process in step S240. Specifically, when generating trajectory point A (the second trajectory point) in path 630, the correct reference point (the first trajectory point) corresponding to trajectory point A is point C. However, if the reference point corresponding to this trajectory point is determined from the reference path 620 solely based on the nearest distance, the determined reference point will be point B. This will cause the planned path 630 to shift towards reference point B, resulting in an unreasonable and infeasible planning result that exceeds the road boundary 610.

[0185] Figure 7 A schematic diagram of a path planning process 700 is shown according to an embodiment of the present disclosure. As shown, the process 700 includes steps S701-S711. Figure 7

[0186] In step S701, perception positioning information is obtained. The perception positioning information can be obtained by various sensors on the vehicle, such as a combination navigation system, a speed sensor, an acceleration sensor, etc. The perception positioning information corresponds to the initial state information above, including a two-dimensional position coordinate in a Cartesian coordinate system, a speed, an acceleration, a heading angle, and a curvature.

[0187] In step S702, a reference line of a target driving behavior (corresponding to the reference path above) is constructed. The reference line can be, for example, a center line of the current lane.

[0188] In step S703, obstacle perception information is obtained, and a detour decision is made for the obstacle based on the obstacle perception information, i.e., whether the obstacle affects the left boundary or the right boundary of the road. The obstacle perception information can be obtained, for example, by inputting the environmental data (such as environmental images, point cloud data, etc.) perceived by the vehicle into a trained object detection model.

[0189] In step S704, an initial boundary of the road is constructed in the Cartesian space according to the reference line. For example, the reference line can be translated to the left and right by half of the width of the current lane to obtain the initial left and right boundaries.

[0190] In step S705, the initial boundary is adjusted according to the obstacle detour direction and the extent to which the vertex of the obstacle intrudes into the initial boundary to obtain a final boundary.

[0191] In step S706, a path optimization problem is constructed, which includes an optimization objective and constraints. The optimization objective can be, for example, to minimize a loss value. The loss value is calculated using a loss function.

[0192] In step S707, reference line, somatic, etc. loss terms in the loss function are constructed. The reference line loss term corresponds to the first loss term above, and the somatic loss term corresponds to the second loss term above.

[0193] In step S708, road boundary, state, etc. constraints are constructed.

[0194] In step S709, the path optimization problem is solved using an iLQR solver. The solving process is further shown in step S710.

[0195] ​In step S710, the solving process includes three sub-problems, i.e., zero control, initial problem and optimization problem. In the zero control sub-problem, an initial solution that is more consistent with the reference line is obtained through PID feedback control according to the expected speed and the reference line. The solving process of the zero control sub-problem can be referred to step S230 in the foregoing. The generated initial solution corresponds to the initial path in the foregoing.

[0196] In the initial sub-problem, the initial solution is optimized once only using the reference line loss term, that is, the loss function only includes the first loss term (cost = cost1), to obtain a starting solution that is more consistent with the reference line.

[0197] In the optimization sub-problem, the road boundary constraint and the road constraint are added, and the on somatic loss term and the speed loss term (corresponding to the third loss term in the foregoing) are added in the loss function, to obtain an optimized solution that can avoid obstacles and meet the vehicle dynamics constraint.

[0198] It should be noted that the zero control sub-problem and the initial sub-problem each include only one optimization, that is, only one path is generated. In the optimization sub-problem, multiple optimized solutions can be generated step by step through iteration. When a preset termination condition is reached, the currently generated optimized solution is determined as the optimal solution (corresponding to the target path in the foregoing).

[0199] In step S711, the target path for which the optimization solving is successful is output. By controlling the vehicle to travel according to the target path, the vehicle can complete the target driving behavior, such as turning.

[0200] According to the embodiments of the present disclosure, the rationality of path planning for complex driving behaviors of an autonomous vehicle can be improved, and the success rate and solving effect of the optimization problem solving are both excellent, and the stability and flexibility are good.

[0201] According to the embodiments of the present disclosure, the optimization problem (including the optimization target and the constraint) is constructed directly in the Cartesian coordinate system and solved by using the iLQR solver, which can ensure the accuracy and speed of the solving. In order to improve the solving effect for complex driving behaviors (such as turning), the zero control scheme is designed, which can provide an initial solution with good effect and strong feasibility based on the PID negative feedback theory, thereby improving the solving speed of the subsequent optimization problem. Further, the form of the loss function is improved, so that the finally generated optimal solution can take into account the safety and comfort of vehicle travel. In view of the particularity of the turning scene, the point finding method of the reference line is improved (i.e., incremental point finding), thereby improving the path planning effect of the turning scene.

[0202] According to the embodiments of the present disclosure, a path planning device is also provided. Figure 8 The structural block diagram of the path planning device 800 according to the embodiments of the present disclosure is shown. As shown in FIG. 8, the path planning device 800 includes a path planning unit 810. Figure 8As shown, the apparatus 800 includes a first determining module 810, a first obtaining module 820, a first generating module 830, and a second generating module 840.

[0203] The first determining module 810 is configured to determine, based on initial state information of a vehicle and obstacle perception information, a target road region in which the vehicle can perform a target driving behavior.

[0204] The first obtaining module 820 is configured to obtain a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior.

[0205] The first generating module 830 is configured to generate, based on the initial state information and the reference path, an initial path for implementing the target driving behavior.

[0206] The second generating module 840 is configured to optimize, based on a preset optimization target, the initial path under constraints of the target road region and the state information range, to generate a target path for implementing the target driving behavior.

[0207] According to embodiments of the present disclosure, first, the initial state information of the vehicle and the reference path are combined to quickly generate an initial path that is close to the reference path and is consistent with the actual state of the vehicle and has strong feasibility. Then, the initial path with strong feasibility is optimized in combination with the current road constraints and behavior state constraints, which can ensure the solving speed and quality of the path optimization, so as to quickly obtain a reasonable and feasible target path, improve the rationality and efficiency of the path planning for complex driving behaviors, and enable the vehicle to safely and efficiently automatically travel.

[0208] According to some embodiments, the initial state information includes position information, and the first determining module includes a first determining unit configured to determine, based on the position information, an initial road region corresponding to the target driving behavior, and a first adjusting unit configured to adjust a boundary of the initial road region based on the obstacle perception information to obtain the target road region.

[0209] According to some embodiments, the obstacle perception information comprises a plurality of vertex coordinates of the obstacle, and the first adjusting unit comprises: a first determining sub-unit, configured to, in response to at least one vertex coordinate in the plurality of vertex coordinates being located within the initial road region, determine, based on the at least one vertex coordinate, a contour line of the obstacle within the initial road region; a second determining sub-unit, configured to determine, as a target boundary segment, a first boundary segment intersecting the contour line and a second boundary segment of a preset length adjacent to the first boundary segment; and an adjusting sub-unit, configured to adjust the target boundary segment towards the inside of the initial road region, so that the adjusted target boundary segment bypasses the contour line.

[0210] According to some embodiments, the first obtaining module is configured to obtain the reference path by any one of the following operations: determining, as the reference path, a feasible path obtained by searching in the target road region using a search algorithm; inputting at least information of the target road region into a trained machine learning model to obtain the reference path output by the machine learning model; or obtaining the reference path marked in a map.

[0211] According to some embodiments, the first generating module is further configured to generate, based on the initial state information and the reference path, the initial path using a preset target PID controller.

[0212] According to some embodiments, the reference path is represented as a first sequence of trajectory point information, the first sequence of trajectory point information comprising first state information of a plurality of first trajectory points respectively, the initial path is represented as a second sequence of trajectory point information, the second sequence of trajectory point information comprising second state information of a plurality of second trajectory points respectively, and the first generating module comprises: a trajectory point planning unit, configured to sequentially perform, for each first trajectory point in the plurality of first trajectory points, the following operations to determine the second state information of the second trajectory point corresponding to the first trajectory point: determining the first state information of the first trajectory point as desired state information; obtaining current state information, wherein, in response to the first trajectory point being a first first trajectory point in the plurality of first trajectory points, the initial state information is determined as the current state information, and in response to the first trajectory point being any first trajectory point other than the first first trajectory point, the second state information of the second trajectory point corresponding to a previous first trajectory point is determined as the current state information; generating, based on a deviation between the current state information and the desired state information, control information using the target PID controller; and determining, based on the control information, the second state information of the second trajectory point corresponding to the first trajectory point.

[0213] According to some embodiments, the trajectory point planning unit comprises: a third determining sub-unit, configured to determine state change information based on the control information; and a superimposing sub-unit, configured to superimpose the current state information and the state change information to obtain second state information of a second trajectory point corresponding to the first trajectory point.

[0214] According to some embodiments, the control information comprises a plurality of control information items, and the state change information comprises a plurality of change information items, each of the plurality of change information items being obtained by integrating at least one control information item of the plurality of control information items.

[0215] According to some embodiments, the apparatus 800 further comprises: a second obtaining module, configured to obtain a plurality of candidate PID controllers; a scoring module, configured to, for any candidate PID controller of the plurality of candidate PID controllers: generate a test initial path based on test initial state information and a test reference path by using the candidate PID controller; and calculate a difference between the test initial path and the test reference path; and a second determining module, configured to determine the candidate PID controller with the smallest difference as the target PID controller.

[0216] According to some embodiments, the optimization target comprises minimizing a loss value, and the second generating module comprises: a loss unit, configured to calculate a loss value of the initial path based on a preset loss function; and a second adjusting unit, configured to adjust the initial path under the constraints of the target road region and the state information range to reduce the loss value, to obtain the target path.

[0217] According to some embodiments, the loss function comprises a first loss item for indicating a difference between the initial path and the reference path, and a second loss item for indicating a smoothness of the initial path.

[0218] It should be understood that, Figure 8 The various modules and units of the apparatus 800 shown in FIG. 8 can correspond to the various steps in the method 200 described with reference to Figure 2 The operations, features and advantages described above for the method 200 apply equally to the apparatus 800 and the modules and units included therein. For the sake of brevity, certain operations, features and advantages are not described again here.

[0219] Although specific functions are discussed above with reference to specific modules, it should be noted that the functions of the various modules discussed herein can be split into multiple modules, and / or at least some of the functions of multiple modules can be combined into a single module.

[0220] It should also be appreciated that various technologies described herein can be described in the general context of software hardware elements or program modules being executed on one or more computing devices. As such, the technology described here can be implemented with various software Figure 8 The various units described can be implemented in hardware or in hardware combined with software and / or firmware. For example, the units can be implemented as computer program code / instructions configured to be executed in one or more processors and stored in a computer-readable storage medium. Alternatively, the units can be implemented as hardware logic / circuitry. For example, in some embodiments, one or more of the modules 810-840 can be implemented together in a System on Chip (SoC). The SoC can include an integrated circuit chip (which includes one or more of a processor (e.g., a Central Processing Unit (CPU), a microcontroller, a microprocessor, a Digital Signal Processor (DSP), etc.), memory, one or more communication interfaces, and / or other circuitry) and can optionally execute received program code and / or include embedded firmware to perform functions.

[0221] According to an aspect of the present disclosure, an electronic device is also provided, comprising: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the path planning method 200.

[0222] According to an aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is also provided, wherein the computer instructions are used to enable the computer to perform the path planning method 200.

[0223] According to an aspect of the present disclosure, a computer program product is also provided, comprising a computer program which, when executed by a processor, implements the path planning method 200.

[0224] According to an aspect of the present disclosure, an autonomous vehicle is also provided, comprising the electronic device.

[0225] Reference is made to Figure 9The present invention describes a structural block diagram of an electronic device 900 that can serve as a server or client of the present disclosure, which is an example of a hardware device that can be applied to various aspects of the present disclosure. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0226] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0227] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, output unit 907, storage unit 908, and communication unit 909. Input unit 906 can be any type of device capable of inputting information to device 900. Input unit 906 can receive input numerical or character information and generate key signal inputs related to user settings and / or function control of the electronic device, and may include, but is not limited to, a mouse, keyboard, touchscreen, trackpad, trackball, joystick, microphone, and / or remote control. Output unit 907 can be any type of device capable of presenting information, and may include, but is not limited to, a monitor, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 908 may include, but is not limited to, a hard disk and an optical disk. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth devices, 802.11 devices, Wi-Fi devices, WiMAX devices, cellular communication devices, and / or the like.

[0228] The computing unit 901 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 901 performs various methods and processes described above, such as the path planning method 200. For example, in some embodiments, the method 200 can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded onto the RAM 903 and executed by the computing unit 901, one or more steps of the method 200 described above can be performed. Alternatively, in other embodiments, the computing unit 901 can be configured to perform the method 200 by any other appropriate means, such as by means of firmware.

[0229] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0230] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces a means for implementing the functions / acts specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0231] In the context of this disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0232] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0233] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

[0234] The computer system can include clients and servers. This relationship can be. The servers are typically remote from the users who use clients to interact with these servers. These clients and servers are often interconnected via communica tion networks. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The servers can be cloud servers, servers of a distributed system, or servers incorporating blockchain.

[0235] It should be understood that the various forms of flow illustrated above can be re-ordered, steps added or removed. For example, the steps recited in the present disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technology disclosed in the present disclosure are achieved, which is not limited herein.

[0236] While embodiments or examples of the present disclosure have been described with reference to the drawings, it should be understood that the methods, systems, and devices described above are merely exemplary embodiments or examples, and the scope of the present disclosure is not limited by these embodiments or examples. Various elements in the embodiments or examples can be omitted or replaced by equivalent elements thereof. Furthermore, the steps can be performed in an order different from that described in the present disclosure. Further, the various elements in the embodiments or examples can be combined in various ways. It is important that many of the elements described herein can be replaced by equivalent elements that appear after the present disclosure as technology evolves.

Claims

1. A path planning method, comprising: determining, based on initial state information of a vehicle and obstacle perception information, a target road region available for the vehicle to perform a target driving behavior; obtaining a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior; generating, based on the initial state information and the reference path, an initial path for implementing the target driving behavior by using a preset target PID controller, the reference path being represented as a first sequence of trajectory point information, the first sequence of trajectory point information including first state information of each of a plurality of first trajectory points, the initial path being represented as a second sequence of trajectory point information, the second sequence of trajectory point information including second state information of each of a plurality of second trajectory points, the generating, based on the initial state information and the reference path, of the initial path for implementing the target driving behavior by using the preset target PID controller comprising: sequentially performing, for each of the plurality of first trajectory points, the following operations to determine the second state information of the second trajectory point corresponding to the first trajectory point: determining the first state information of the first trajectory point as desired state information; obtaining current state information, wherein, in response to the first trajectory point being a first one of the plurality of first trajectory points, the initial state information is determined as the current state information, and in response to the first trajectory point being any one of the first trajectory points other than the first one, the second state information of the second trajectory point corresponding to a previous first trajectory point is determined as the current state information; generating, based on a deviation between the current state information and the desired state information, control information by using the target PID controller; and determining, based on the control information, the second state information of the second trajectory point corresponding to the first trajectory point; and optimizing, under constraints of the target road region and the state information range, the initial path based on a preset optimization objective to generate a target path for implementing the target driving behavior. The initial state information includes position information, and the determining, based on the initial state information and the obstacle perception information, of the target road region available for the vehicle to perform the target driving behavior comprises:

2. The method of claim 1, wherein, determining, based on the position information, an initial road region corresponding to the target driving behavior; and adjusting, based on the obstacle perception information, a boundary of the initial road region to obtain the target road region. The obstacle perception information includes a plurality of vertex coordinates of an obstacle, and the adjusting, based on the obstacle perception information, of the boundary of the initial road region to obtain the target road region comprises:

3. The method of claim 2, wherein, in response to at least one of the plurality of vertex coordinates being located within the initial road region, determining, based on the at least one vertex coordinate, a contour line of the obstacle within the initial road region. ​ determine a first boundary segment intersecting the contour line and a second boundary segment of a preset length adjacent to the first boundary segment as a target boundary segment; and adjust the target boundary segment towards the inside of the initial road region, so that the adjusted target boundary segment bypasses the contour line.

4. The method of any one of claims 1-3, wherein, The obtaining of the reference path for implementing the target driving behavior includes any of the following operations: determining a feasible path obtained by searching in the target road region by using a search algorithm as the reference path; inputting at least information of the target road region into a trained machine learning model to obtain the reference path output by the machine learning model; or obtaining the reference path marked in a map.

5. The method of claim 1, wherein, The determining of the second state information of the second trajectory point corresponding to the first trajectory point based on the control information includes: determining state change information based on the control information; and superimposing the current state information and the state change information to obtain the second state information of the second trajectory point corresponding to the first trajectory point.

6. The method of claim 5, wherein, The control information includes a plurality of control information items, and the state change information includes a plurality of change information items, each change information item being obtained by integrating at least one control information item in the plurality of control information items.

7. The method of claim 1, further comprising: obtaining a plurality of candidate PID controllers; for any candidate PID controller in the plurality of candidate PID controllers: generating a test initial path by using the candidate PID controller based on test initial state information and a test reference path; and calculating a difference between the test initial path and the test reference path; and determining the candidate PID controller with the smallest difference as the target PID controller. The optimization target includes minimizing a loss value, and the optimizing of the initial path based on a preset optimization target under the constraints of the target road region and the state information range includes: calculating a loss value of the initial path based on a preset loss function; and 8. The method of claim 1, wherein, adjusting the initial path under the constraints of the target road region and the state information range to reduce the loss value, to obtain the target path. The loss function includes a first loss term for indicating a difference between the initial path and the reference path and a second loss term for indicating a smoothness of the initial path.

10. A path planning apparatus, comprising:

9. The method of claim 8, wherein, a first determining module configured to determine a target road region available for a vehicle to perform a target driving behavior based on initial state information of the vehicle and obstacle perception information; a first obtaining module configured to obtain a reference path for implementing the target driving behavior and a state information range corresponding to the target driving behavior, the state information range indicating a value range of state information of the vehicle when performing the target driving behavior; ​ ​ The first generation module is configured to generate, based on the initial state information and the reference path, an initial path for realizing the target driving behavior by using a preset target PID controller, the reference path being represented as a first sequence of trajectory point information, the first sequence of trajectory point information including first state information of each of a plurality of first trajectory points, the initial path being represented as a second sequence of trajectory point information, the second sequence of trajectory point information including second state information of each of a plurality of second trajectory points, and the first generation module including: a trajectory point planning unit configured to sequentially perform the following operations on each of the plurality of first trajectory points to determine the second state information of the second trajectory point corresponding to each of the plurality of first trajectory points: determining the first state information of the first trajectory point as expected state information; obtaining current state information, wherein, in response to the first trajectory point being a first first trajectory point in the plurality of first trajectory points, the initial state information is determined as the current state information, and in response to the first trajectory point being any first trajectory point other than the first first trajectory point, the second state information of the second trajectory point corresponding to a previous first trajectory point is determined as the current state information; generating control information by using the target PID controller based on a deviation between the current state information and the expected state information; and determining the second state information of the second trajectory point corresponding to the first trajectory point based on the control information; and a second generation module configured to optimize the initial path based on a preset optimization target under the constraints of the target road region and the state information range to generate a target path for realizing the target driving behavior. The initial state information includes position information, and the first determination module includes:

11. The apparatus of claim 10, wherein, a first determination unit configured to determine an initial road region corresponding to the target driving behavior based on the position information; and a first adjustment unit configured to adjust a boundary of the initial road region based on the obstacle perception information to obtain the target road region. The obstacle perception information includes a plurality of vertex coordinates of an obstacle, and the first adjustment unit includes:

12. The apparatus of claim 11, wherein, a first determination subunit configured to, in response to at least one vertex coordinate in the plurality of vertex coordinates being located within the initial road region, determine, based on the at least one vertex coordinate, a contour line of the obstacle within the initial road region; a second determination subunit configured to determine, as a target boundary segment, a first boundary segment intersecting the contour line and a second boundary segment of a preset length adjacent to the first boundary segment; and an adjustment subunit configured to adjust the target boundary segment towards the inside of the initial road region so that the adjusted target boundary segment bypasses the contour line. The first obtaining module is configured to obtain the reference path by any of the following operations:

13. The apparatus of any of claims 10-12, wherein, determining, as the reference path, a feasible path obtained by searching in the target road region by using a search algorithm; ​ inputting at least information of the target road region into a trained machine learning model to obtain the reference path output by the machine learning model; or acquiring the reference path marked in a map.

14. The apparatus of claim 10, wherein, The trajectory point planning unit comprises: a third determination sub-unit configured to determine state change information based on the control information; and a superposition sub-unit configured to superimpose the current state information and the state change information to obtain second state information of a second trajectory point corresponding to the first trajectory point.

15. The apparatus of claim 14, wherein, The control information comprises a plurality of control information items, and the state change information comprises a plurality of change information items, each of the plurality of change information items being obtained by integrating at least one control information item in the plurality of control information items.

16. The apparatus of claim 10, further comprising: a second acquisition module configured to acquire a plurality of candidate PID controllers; a scoring module configured to, for any candidate PID controller in the plurality of candidate PID controllers: generate a test initial path based on test initial state information and a test reference path using the candidate PID controller; and calculate a difference between the test initial path and the test reference path; and a second determination module configured to determine the candidate PID controller with the smallest difference as the target PID controller. The optimization target comprises minimizing a loss value, and the second generation module comprises: a loss unit configured to calculate a loss value of the initial path based on a preset loss function; and 17. The apparatus of claim 10, wherein, a second adjustment unit configured to adjust the initial path under constraints of the target road region and the state information range to reduce the loss value, to obtain the target path. The loss function comprises a first loss item for indicating a difference between the initial path and the reference path, and a second loss item for indicating a smoothness of the initial path.

19. An electronic device comprising:

18. The apparatus of claim 17, wherein, at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9. The computer instructions are used to enable a computer to perform the method of any one of claims 1-9. The computer program, when executed by a processor, implements the method of any one of claims 1-9.

20. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, 22. An autonomous vehicle comprising the electronic device of claim 19.

21. A computer program product comprising a computer program, wherein, ​ ​

Images